Modifications of Optimum Adaptation Zones for Soybean Maturity Groups in the USA

Search PMN

L. X. Zhang, Delta Research and Extension Center, Mississippi State University, Stoneville 38776; S. Kyei-Boahen, International Institute of Tropical Agriculture (IITA), c/o IIAM-PAN, Nampula, Mozambique; J. Zhang, Tobaco Research Institute, Chinese Academy of Agricultural Sciences, Qingdao 266101; M. H. Zhang, Department of Land, Air and Water Resources, University of California, Davis 95616; T. B. Freeland, Ayres Delta Implement Inc., Leland, MS 38756; C. E. Watson, Jr., Oklahoma Agricultural Experiment Station, Oklahoma State University, Stillwater 74078; and Xingmei Liu, Zhejiang University, Zhejiang, China 310029

Abstract

Soybean cultivars are generally adapted within a narrow north-south geographical zone for full-season growth due primarily to photoperiod response. The areas of adaptation were empirically defined more than three decades ago and re-examination is needed. The accuracy of adaptation area determination can be improved by utilizing updated information, including changes in production practices and new technologies. The objective of this paper was to use current soybean yield data from experiments conducted across soybean producing states in the US to develop optimum zones of adaptation for soybean maturity groups (MGs) in the continental US. Data from state soybean variety trials conducted in 1998-2003 were obtained from 139 locations, and were used to create regional adaptation map using ArcGIS. The MG 0 cultivars are adapted best to the region north of latitude 46°N, whereas succeeding groups are adapted further south. Each of the MGs 0 to III is adapted best within approximately two degrees latitude covering an area equivalent to 220 km wide from north to south. The zones of adaptation for MGs IV, V, and VI are wider than those for the earlier maturing cultivars. Overall, the regions of adaptation for the early-maturing cultivars (MG 0 to III) have not changed; however, the adapted zones for MGs IV, V, and VI are much broader than previously thought. Groups VII and VIII, which dominated production areas in the South decades ago are now planted on a limited basis.

Introduction

Time of flowering and maturity in soybean are important agronomic characters that determine its geographical adaptation. Based on these traits, soybean has been classified into 13 maturity groups (MG). In North America, soybean cultivars with 000, 00, and 0 groups are the earliest in maturity and are mainly adapted to production areas in southern Canada (11). Maturity groups I and II are typically grown in the northern region of the US, whereas succeeding groups are grown further south. According to Hartwig (18), MG VIII is the latest grown cultivar in the continental US, but MGs IX and X, which flower later than MG VIII when grown in the South, have been classified. This pattern of adaptation from north to south is due primarily to differences in sensitivity to photoperiod (10,30). The MG of a particular soybean cultivar is assigned with a numeric number (often up to one decimal point) based on the days from planting to maturity at a defined latitude (or day length) and a specific planting date under optimum environmental conditions. For example, if a soybean cultivar is planted on April 15th at Stoneville (latitude 33.4°N), Mississippi (or similar latitude), and matures in about 130 to 135 days, this cultivar be classified as MG 4.4 (40).

Vast information has accumulated over the years since the report of Garner and Allard (16) on the effect of temperature and photoperiod, and the interaction of these factors on the time of flowering in soybean. In most studies, warm spring temperatures promoted earlier flowering, however cultivars differed considerably in flowering response to photoperiod (12,22,36). For this reason, Hartwig (19) concluded that photoperiodism is more important in determining the area of adaptation for soybean than any other major crop. The cultivars selected for the southern latitudes generally respond to shorter days than the cultivars adapted to the northern region; thus the southern cultivars would flower too late in the north whereas the northern cultivars would flower too early in the South (20,30).

Scott and Aldrich (30) defined hypothetical zones of adaptation for 10 soybean MGs in the US (Fig. 1) and has been the most widely referenced document regarding the area of adaptation for specific soybean maturity groups. The zones of adaptation were developed mostly based on empirical data available at the time. Research information accumulated during the past three decades, in addition to current technology and Geographical Information Systems (GIS) provide an opportunity to re-examine these zones of adaptation to guide soybean production and build better soybean production models.

Fig. 1. Map showing regions in the continental USA and southern Canada where various soybean maturity groups (MG) are adapted for full-season growth. The MG designated 00 are the earliest maturing, and VIII the latest maturing. The higher the number the further south the MG is adapted for full-season growth. Adapted from Scott and Aldrich (30).

To develop an accurate and reliable representation of the various MGs at the respective areas of adaptation, information on optimum planting, harvest dates, and agronomic performance, such as seed yield at various locations, are required. Tremendous information is readily available from state variety trials and could be helpful in this endeavor. In most of the soybean producing states, official variety trials are important component of the soybean research programs. Most of these experiments are conducted under optimum planting conditions using the best management practices for soybean production in the region. Thus, data obtained from these studies are reliable for production and research purposes. The objectives of this study were to analyze the distribution of soybean MGs adapted to the continental US and develop regions of adaptation based on current information and technology.

Methodology for Developing Optimum Zones for Soybean MGs

Data source. Data used in this study were obtained from the websites of 28 soybean-producing states (Table 1) compiled from individual state variety trials. The trials were conducted across many locations and under irrigated and non-irrigated conditions from 1998-2003. The experiments consisted of different soybean genotypes with various levels of disease resistance, and included roundup ready and conventional cultivars.

Table 1. List of region, latitude, and planting period for each state where the soybean yield data were obtained. Corresponded planting period indicates the range of acceptable (including optimum) planting dates in that region. Variety trial data were selected within the limit of the correspondent period. Letters in parenthesis indicate the region of the state included (n = North, c = Central, and s = South).

Region	Latitude	Planting period	States included
North	> 46°	5/15–6/25	ND, MN (n)
North-central	46°-42°	5/05–6/20	SD, MN (s), IA (n), WI, MI, NE (n), NY
Central	42°-38°	4/25–6/15	OK, IA (s), NE (s), IL, KS (n), MO(n), OH, IN, PA, WA, VA (n), NJ, DE, MD
South-central	38°-34°	4/15–6/10	KS (s), MO (s), KY, AK (n), TN, MS (n), AL (n), GA (n), NC, SC (n)
South	34°-30°	4/05–6/05	AK (s), LA, MS (c, s), AL (c, s), GA (c, s), SC (s), TX

Criteria for data selection. The geographical adaptation of soybean is based on its genetic makeup. However, environmental factors, such as photoperiod, temperature, and some other factors may also influence soybean production and final yields. To better describe soybean adaptation, some criteria were developed to ensure that the data were obtained under optimum conditions.

Irrigation vs. non-irrigation. Initial analysis of the data from irrigated trials in all the soybean producing states in recent years (1998-2003) indicated that over 90% of the cultivars yielded at least 3030 kg/ha under normal conditions; hence, soybeans with yield lower than this value were presumably not grown under optimum field conditions, and data were not included in the study. Non-irrigated soybeans may receive sufficient precipitation in some years and could produce yields similar to irrigated soybeans. Thus, data from non-irrigated fields that met this criterion were included. Most trials from Midwest and early planted trials in the Midsouth had high yields, which indicated that the trials were under optimum production system for the regions. In the Southeast soybeans used were mostly late MGs and planted late under non-irrigated conditions (mostly due to economic reasons); resulting in relatively lower yields. Due to the nature of this study, which was to discuss the optimum adaptation zones for soybean MG, few available data were included for this study. For the similar reason, data from part of the Southwest, specifically in Texas, were not included due to low yield under non-optimum production system.

Planting period. Geographical location and associated climate conditions restrict planting time to a particular period during the growing season. Thus, the optimum planting date for a particular location may be different from another location, in particular when the latitudinal difference is large. The potential yield of a given maturity groups would not be reflected accurately when planted before or after the optimum planting period. For this reason, the continental US was hypothetically divided into five zones based on the latitude: North, 46°N or above, then progressively south at an interval of 4° of latitude for each zone (North-central, Central, South-central, and South). Based on the above zones, over 90% of the trials of the individual states have fall into planting periods shown in Table 1. Hence, variety trials conducted before or after these dates were not considered.

Determination of the Optimum Maturity Group Value at a Location

The data indicated that the various soybean MGs were not equally represented in the individual experiments. When other conditions are held the same, a cultivar with the most optimum MG (most adapted) should produce the highest yield. In order to determine the optimum MG in an adapted zone more accurately, yield is considered as the main weighing factor; therefore, a weighted mean method was used to calculate the optimum MG at a particular location to restore balance (32). The maturity group value (V_mg) was calculated using the following equation:

where w is the yield, and x is the value of maturity group.

In this study, the values of MG I, II, … and VIII were defined from 1.0 to 1.9, 2.0 to 2.9, … and 8.0 to 8.9, respectively, as they were conventionally suggested. However, the values of MG 0, 00, and 000, were specifically defined from 0.1 to 0.9, 0.01 to 0.09, and 0.001 to 0.009, respectively, to avoid unconventional numbers of 0.0, 0.00, and 0.000, which may cause confusion during the calculations.

Map Development

Data from 139 locations met the criteria set above and were used for the development of soybean MG adaptive zones in the US using ArcGIS 8.2 (14). The data were compiled and exported from Excel into a text format capable of being utilized by ArcGIS. A shapefile was created utilizing the latitude and longitude data for each location inside ArcCatalog. The shape-file was then added to a project in ArcMap that contained the shapefile of the continental US as a background reference. The data points were displayed (Fig. 2). To create adaptation areas suited for soybean growth based on data obtained from these locations, an interpolation between the data was completed using three approaches: the spline and inverse distance weighted (IDW) methods using ArcGIS, and the kriging method utilizing ArcMap and the Spatial Analyst extension.

Fig. 2. Map showing locations in the continental USA where the soybean yield trials were conducted and from where the yield data were obtained.

Spline interpolation uses polynomial interpolation, and the polynomial of degree n which interpolates the data set is uniquely defined by the data points. The spline of degree n which interpolates the same data set is not uniquely defined, and we have to fill in n-1 additional degrees of freedom to construct a unique spline interpolant. For example, given n+1 distinct knots x_i such that

with n+1 knot values y_i we are trying to find a spline function of degree n

where each S_i(x) is a polynomial of degree k.

Linear spline interpolation is the simplest form of spline interpolation and is equivalent to linear interpolation. The data points are graphically connected by straight lines. The resultant spline is just a polygon. Algebraically, each S_i is a linear function constructed as

The spline must be continuous at each data point, that is

Inverse distance weighted technique is to interpolate scattered data points, and the method is based on the assumption that the interpolating surface should be influenced most by the nearby points and less by the more distant points. The interpolating surface is a weighted average of the scattered data points and the weight assigned to each scattered data point diminishes as the distance from the interpolation point to the scattered data point increases.

The simplest form of inverse distance weighted interpolation is sometimes called "Shepard's method" (31). The equation used is as follows:

where n is the number of scatter points in the set, f_i are the prescribed function values at the scatter points (e.g., the data set values), and w_i are the weight functions assigned to each scatter point.

Ordinary Kriging estimates the unknown value using weighted linear combinations of the available data points:

The error of i-th estimate, r_i, is the difference of estimated value and true value at that same location:

The Spline and the IDW methods are referred to as deterministic interpolation methods because they are directly based on the surrounding measured values or on specified mathematical formulas that determine the smoothness of the resulting surface. Kriging method is based on statistical models that include autocorrelation, that is, the statistical relationship among the measured points. After considering the error estimates and the assumptions of each method (25), the Kriging method was selected. The Kriging method also produced the smoothest interpolation and did not extend the interpolation beyond the scope of the data; it also gave the best visual appearance to the interpolated data.

After the first interpolation, it was evident that the data included some outliers, as "bull’s-eyes" were included in several areas of the interpolation. Four data points were discarded and the interpolation was run on the remaining dataset. Figure 3 shows the raster created by the Kriging interpolation method with an output cell size of 0.05 decimal degrees, and a 12-point variable radius search on a spherical semi-variogram model. A mask shapefile was created in ArcCatalog that covered the states where the data were collected. The mask was then split based on the interpolation into seven areas that followed the maturity group delineations. This mask was then cut to extend only to the states whose data were included in the data set (Fig. 3). Due to the statistical non-significance on the variations between the kriged maturity group contour lines and the splined smoothing maturity groups (24), we combined the results from both Kriging and spline interpolations to produce the improved adaptation zones for soybean maturity groups (Fig. 4).

Fig. 3. Map showing interpolated data for areas of adaptation for soybean maturity groups in the continental USA.

Fig. 4. Map showing the zones of adaptation for soybean maturity groups in the continental USA.

Optimum Adaptation Zones for Soybean Maturity Groups

The locations where soybean yield data were selected are indicated in Fig. 2. The number of locations on the map varies from state to state due to the differences in the number of trials conducted in the individual states. In addition, data from locations that did not meet the criteria for data selection (including yield threshold and optimum planting date) were not considered. For example, eight or more locations in Nebraska, Missouri, Illinois, and Wisconsin were considered, whereas only one or two locations in Georgia, Pennsylvania, and Oklahoma were selected. No data were available or considered in Texas, West Virginia, and Florida. The zones of adaptation for soybean MGs based on the data selected from the various states are shown in Fig. 4. A zone of adaptation in this study represents a defined region where a given cultivar of a particular MG is best adapted and produces the highest seed yield when planted within the optimum planting dates (Table 1), under irrigation or in a year with adequate rainfall, in the absence or low incidence of pests and diseases and when the best management system is practiced. It should be emphasized that this in no way implies that the adapted soybean MG can only be grown successfully in that particular region. Typically, cultivars from three adjacent MGs are often grown successfully at a specific site within a particular MG adaptation zone, especially in the southern region where planting date can be relatively flexible.

The data indicate that MG 0 (or earlier) cultivars are adapted to the region north of latitude 46°N comprising North Dakota, the northern half of Minnesota, and areas north of Lake Michigan. The zone across South Dakota, the southern half of Minnesota, and northern Wisconsin are best adapted to MG I, whereas areas from Nebraska extending across Iowa, southern Wisconsin, northern Illinois, and southern Michigan are suited to MG II. The group III cultivars have their highest yield potential in greater part of Kansas, Missouri, Illinois, and Ohio extending to Pennsylvania: mostly within latitude 38°N and 41°N. The MG IV cultivars are mainly adapted to the zone from Oklahoma across Kentucky and Tennessee to the east coast, whereas the highest yield potential for the MG V cultivars can be found mostly in the region extending from southern Arkansas, Louisiana, and Mississippi to North Carolina. The most adapted late-maturing cultivars were found to be those from MG VI, which are used primarily in southern Alabama, and near the Atlantic coast of Georgia and South Carolina.

The pattern of soybean MG adaptation indicated in Fig. 4 is due primarily to the sensitivity to day length (10). Soybean is a short-day plant where flowering occurs only when the day length is shorter than the critical photoperiod. Flowering in cultivars adapted to the southern USA (lower latitudes) are typically delayed by long days when grown in the northern USA (higher latitudes) making them too late for full-season growth. In contrast, the cultivars adapted to the northern USA flower too early when grown in the southern USA where days are much shorter. In this environment, the plants do not develop sufficient vegetative biomass for optimal yield (1,4,5). Sensitivity to photoperiod can be modulated by temperature (35,37). In turn, photoperiod and temperature interact with genotype to control soybean growth and development, from germination through the onset of flowering to maturity (9,20,22,33).

Many studies indicated that when the photoperiod is less than the critical value, higher temperatures result in early flowering whereas cool temperatures delay flowering (2,3,12,22,33,35). It is now known that six loci, each with two alleles [dominant (E) and recessive (e)] control days to flowering and maturity in soybean (7,8,11,23,26). These reports have provided a better understanding of the genetic basis of adaptation of soybean genotypes to different photoperiodic regimes, which vary with latitude and growing season. The dominant alleles have been shown to delay flowering in photoperiods longer than the critical value; the magnitude of the delay depending on the number of dominant alleles (9,10,35,36). Furthermore, the individual dominant alleles delay flowering in long days to various extent, and the same could be said about the combinations among the photoperiod-sensitivity genes (28,35).

Roberts et al. (27,28) argued that the more photoperiod sensitive gene combinations are well suited to the short but warm days of the tropics where flowering would otherwise be premature, whereas the less sensitive combinations are appropriate to high latitudes where photoperiods are long during the growing season. Hence, the photoperiod gene combination e₁e₂e₃ is typical of MG I, and perhaps other earlier maturing cultivars adapted to latitudes 44°N and higher in North America (38). There are evidence that the gene combinations of e₁E₂E₃, E₁e₂E₃, E₁E₂e₃, and E₁e₂e₃ are typical of MG IV and V, whereas E₁E₂E₃ is typical of MG V and above (38) and would normally be cultivars adapted to latitudes below 36°N in North America (35). This information agrees with the data presented in Fig. 4, although almost half of the region adapted to MG IV is above latitude 36°N. Similarly, the geographical adaptation zones for early maturing soybean cultivars from MG III and lower shown in the present paper (Fig. 4) are to a large extent consistent with those presented by Scott and Aldrich (30) in Fig. 1.

The region of adaptation for each of the MGs 0 to III is within approximately two degrees latitude and equivalent to a distance of 220 km from north to south, and is much narrower than those for MGs IV and V. This is in agreement with the suggestions by Hartwig (19) and Scott and Aldrich (30). However, there are four major differences in the zones of adaptation between Figs. 4 and 1 regarding MG IV and the later maturing MGs:

(i) The zones mostly adapted to MG IV and V are much broader than that indicated by Scott and Aldrich (30);

(ii) The zone of adaptation for MG IV in Fig. 4 occupies the bands adapted to MG IV, V and part of the area adapted to VI shown in Fig. 1;

(iii) The area adapted to MG V in Fig. 4 cover part of the region of adaptation for MG VI, the whole area for VII and portion for VIII in Fig. 1; and

(iv) The latest MG most adapted to the continental US was found to be MG VI, and cover an area within the region previously earmarked for MG VIII by Scott and Aldrich (30).

The main reason for the similarities between Figs. 1 and 4 at latitudes higher than 40°N is that soybean production practices, in particular planting date and MG cultivars grown have not changed much since the last three decades. In these regions, the weather often restricts soybean planting until after 1 May or mid-May when soil temperatures become warm enough to facilitate germination. This partly accounts for the narrower bands of adaptation for MGs 0 to III relative to the later maturing MGs as indicated in Fig. 4. However, production trends have undergone a number of changes in the southern US, especially the Mid-South, in recent years. The most significant changes include the shift in planting date from late May-June to early planting, typically early April to early May using relatively early MG cultivars particularly MG IV and V.

Late-maturing cultivars from MG VI, VII, and VIII were dominant genotypes grown in the regions two decades ago. The reason being that late planting, between late May and late June or even early July, was recommended for soybeans used then (18,19,17). According to Zhang et al. (40), at Stoneville location in Mississippi (33.4°N, 90.9°W), if an early soybean cultivar, such as a late MG IV, is planted on 1 June, it may reach about V2 (15) stage around 21 June, when day length is at its maximum. Day length declines progressively after late June, causing soybean flowering around early to mid-July. Thereafter, plant vegetative growth slows down and whole reproductive growth is under drastically reduced day-length, causing reduction in pod set and seed-fill durations. With this late planting period, early soybeans could not reach their full yield potential. Thus, the plants are shorter and smaller in size and may have inadequate vegetative biomass to produce optimum yield. In contrast, the MGs VI, VII, and VIII grow taller and may achieve adequate vegetative growth and a higher yield advantage when planted late in the South when compared with MGs IV and V. A late MG VI cultivar planted in early June will not flower until 6 to 7 weeks after emergence.

The original rationale for early soybean production system in the South was to avoid drought stress during the reproductive development stage by planting early using early cultivars (6,13,20,21,29,34). However, the effect of photoperiod regimes associated with planting dates on crop phenology has been recognized as one of the major causes of yield increase on irrigated fields (39). With earlier planting dates, such as early April to early May, MG IV or V cultivar can develop sufficient vegetative growth before and after flowering, due to indeterminate growth habits (29) and also have a longer period for reproductive development. For example, the duration between flowering (R1) and full seed stage (R6) for a late MG IV soybean planted on 20 April is 5 to 6 days longer than that for late MG VI planted on 1 June (40). Thus, the yield potentials of the MG IV and V are much higher in the southern latitudes when planted early than previously recognized.

Theoretically, MG V soybeans should be best adapted to the region around the southeastern coast, similar to those in the Mid-South. However, economically, later MG (MG VI) is favored in the production practices due primarily to the availability of late season precipitation during September and early October in addition to the relatively warm conditions during this period in the area.

In conclusion, the areas of adaptation for soybean MGs from 0 to III have not undergone major changes since three decades ago. However, significant changes have occurred with respect to the areas defined for groups IV, V, and VI. The region of adaptation for each of these relatively late maturing groups is much wider than previously reported and cover approximately 3 to 5° latitude compared with approximately 2° latitude for each of the earlier maturing groups. Groups VII and VIII, which were the dominant cultivars used decades ago are no longer planted commercially in the Mid-South but are planted on a limited basis in the southern US. These changes are mainly due to the adoption of early planting using relatively early-maturing cultivars in the southern latitudes to match crop phenology to water supply during the growing season.

Acknowledgments

The authors acknowledge the work of all individuals involved in the state soybean variety trials that made this study possible. We also express our appreciation to Ms. Juliana Dong for her assistance in data management.

Literature Cited

1. Ball, R. A., Purcell, L. C., and Vories, E. D. 2000. Optimum soybean plant population for a short-season production system in the southern USA. Crop Sci. 40:757-764.

2. Bernard, R. L. 1971. Two major genes for time of flowering and maturity in soybean. Crop Sci. 11:242-244.

3. Board, J. E., and Hall, W. 1984. Premature flowering in soybean yield reductions at non-optimal planting dates as influenced by temperature and photoperiod. Agron. J. 76:700-704.

4. Board, J. E., and Settimi, J. R. 1986. Photoperiod effect before and after flowering on branch development in determinate soybean. Agron. J. 78:995-1002.

5. Boerma, H. R., and Ashley, D. A. 1982. Irrigation, row spacing, and genotype effects on late and ultra-late planted soybeans. Agron. J. 74:995-998.

6. Bowers, G. R. 1995. An early soybean production system for drought avoidance. J. Prod. Agric. 8:112-119.

7. Buzzell, R. I. 1971. Inheritance of a soybean flowering response to fluorescent-daylength conditions. Can. J. Genet. Cytol. 13:703-707.

8. Buzzell, R. I., and Voldeng, H. D. 1980. Inheritance of insensitivity to long daylength. Soybean Genet. Newsl. 7:26-29.

9. Cober, E. R., Stewart, D. W., and Voldeng, H. D. 2001. Photoperiod and temperature responses in early-maturing, near-isogenic soybean lines. Crop Sci. 41:721-727.

10. Cober, E. R., Tanner, J. W., and Voldeng, H. D. 1996. Genetic control of photoperiod response I early-maturing, near-isogenic soybean lines. Crop Sci. 36:601-605.

11. Cober, E. R., and Voldeng, H. D. 2001. A new soybean maturity and photoperiod-sensitive locus linked to E1 and T. Crop Sci. 41:698-701.

12. Cooper, R. L. 2003. A delay flowering barrier to higher soybean yields. Field Crops Res. 82:27-35.

13. Edwards, J. T., Purcell, L. C., Vories, E. D., Shannon, J. G., and Ashlock, L. O. 2003. Short-season soybean cultivars have similar yields with less irrigation than longer-season cultivars. Online. Crop Management doi:10.1094/CM-2003-0922-01-RS.

14. ESRI. 2000. Arc/Info user’s guide. Online document Version 8.2. Environmental Systems Research Initiative, Redlands, CA.

15. Fehr, W. R., and Caviness, C. E. 1977. Stages of soybean development. Spec. Rep. 80., Iowa Agric. Exp. Stn., Ames, IA.

16. Garner, W. W., and Allard, H. A. 1930. Photoperiodic response of soybeans in relation to temperature and other environmental factors. J. Agric. Res. 41:719-735.

17. Hartwig, E. E. 1968. Soybean varieties for the Delta area. Mississippi Agric. Exp. Stn. Info. Sheet, 1008, 2, Mississippi State, MS.

18. Hartwig, E. E. 1970. Growth and reproductive characteristics of soybeans [Glycine max (L.) Merr.] grown under short-day conditions. Trop. Sci. 12:47-53.

19. Hartwig, E. E. 1973. Varietal development. Pages 187-210 in: Soybeans: Improvement, Production and Uses. Agron. Monogr. 16, 1st Ed. B. E. Caldwell, ed. ASA, CSSA, and SSSA, Madison, WI.

20. Heatherly, L. G., and Elmore, R. W. 2004. Managing inputs for peak production. Pages 451-536 in: Soybeans: Improvement, Production and Uses. Agron. Monogr. 16, 1st Ed. B. E. Caldwell, ed. ASA, CSSA, and SSSA, Madison, WI.

21. Heitholt, J. J., Farr, J. B., and Sutton, R. L. 2005. Risk management in north Texas soybean: Mid-March soybean plantings uncertain; maturity group IV cultivars reliable. Online. Crop Management doi:10.1094/CM-2005-0329-01-RS.

22. Major, D. L., Johnson, D. J., Tanner, J. W., and Anderson, I. C. 1975. Effects of day length and temperature on soybean development. Crop Sci. 15:174-179.

23. McBlain, B. A., and Bernard, R. L. 1987. A new gene affecting the time of flowering and maturity in soybeans. J. Hered. 78:160-162.

24. Mitas, L., and Mitasova, H. 1988. General Variational Approach to the Interpolation Problem. Comput. Math. Applic. Vol. 16. No. 12. pp. 983–992. Great Britain.

25. Philip, G. M., and Watson, D. F. 1982. A precise method for determining contoured surfaces. Australian Petroleum Exploration Assoc. J. 22:205-212.

26. Piper, E. I., Smit, M. A., Boote, K. J., and Jones, J. W. 1996. The role of minimum temperature in modulating the development rate to flowering in soybean. Field Crops Res. 47:211-220.

27. Roberts, E. H., Summerfield, R. J., Ellis, R. H., and Qi, A. 1993. Adaptation of flowering in crops to climate. Outlook on Agric. 22:105-110.

28. Roberts, E. H., Qi, A., Ellis, R. H., Summerfield, R. J., Lawn, R. J., and Shanmugasundaram, S. 1996. Use of field observations to characterize genotypic flowering responses to photoperiod and temperature: A soybean exemplar. Theor. Appl. Genet. 93:519-533.

29. Savoy, B. R., Cothren, J. T., and Shumway, C. R. 1992. Early-season production systems utilizing indeterminate soybean. Agron. J. 84:394-398.

30. Scott, W. O., and Aldrich, S. R. 1970. Modern Soybean Production. 1st Ed. S & A Publ. Inc, Champaign, IL.

31. Shepard, D. 1968. A two-dimensional interpolation function for irregularly spaced data. Pages 517-524 in: Proc. of the 23rd National Conf. of the ACM. ACM Press, New York.

32. Steel, R. G. D., Torrie, J. H., and Dickey, D. A. 1997. Principles and procedures of statistics: A biometrical approach. 3rd ed. McGraw-Hill, New York.

33. Summerfield, R. J., Lawn, R. J., Qi, A., Ellis, R. H., Roberts, E. H., Chay, P. M., Brouwer, J. B., Rose, J. L., Shanmugasundaram, S., Yeates, S. J., and Sandover, S. 1993. Towards the reliable prediction of time of flowering in six annual crops. II. Soybean (Glycine max). Exp. Agric. 29:253-289.

34. Sweeney, D. W., Granade, G. V., and Burton, R. O., Jr. 1995. Early and traditionally maturing soybean varieties grown in two planting systems. J. Prod. Agric. 8:373-379.

35. Upadhyay, A. P., Ellis, R. H., Summerfield, R. J., Roberts, E. H., and Qi, A. 1994. Characterization of photothermal flowering responses in maturity isolines of soyabean [Glycine max (L.) Merrill] cv. Clark. Ann. Bot. (London) 74:87-96.

36. Upadhyay, A. P., Summerfield, R. J., Ellis, R. H., Roberts, E. H., and Qi, A. 1994. Variation in the duration of the photoperiod-sensitive phases of development to flowering among eight maturity isolines of soyabean [Glycine max (L.) Merrill]. Ann. Bot. (London) 74:97-101.

37. Wallace, D. H., and Yan, W. 1998. Plant breeding and whole crop systems, improving adaptation, maturity, and yield. CAB International, Wallingford, UK.

38. Wang, J., McBlain, B. A., Hesketh, J. D., Woolley, J. T., and Bernard, R. L. 1987. A data base for predicting soybean phenology. Biotronics 16:25-38.

39. Zhang, L. X. 2004. Exploring the feasibility of growing maturity group III soybeans in the Mid-South. Res. Sheet. 1343, Mississippi Agric. For. Exp. Stn., Mississippi State, MS.

40. Zhang, L., Zhang, J., Watson, C. E., and Kyei-Boahen, S. 2004. Developing phenological prediction tables for soybean. Online. Crop Management doi:10.1094/CM-2004-1025-01-RS.